Many academic institutions and Cooperative Groups maintain extensive archives of tumor tissue from patients with matching treatment and clinical outcome information. Such archives have been utilized for molecular characterizations of specific tumors, as well as prognostic studies aimed at investigating the effect of acquired genetic alterations on clinical outcome measures. There is increasing interest in utilizing these tumor archives in pharmacogenetic studies, which are concerned with assessing the associations between constitutional genetic polymorphisms and treatment-related toxicity and prognosis. Little effort has focused on the appropriateness of using diseased tissue as a source of genomic DNA in pharmacogenetic studies. We believe that an important first step in pharmacogenetic studies that utilize stored tumor tissue as a source of genomic DNA should be to demonstrate concordance between polymorphism measured in diseased and paired non-diseased tissue. Since our long-term interest lies in conducting a pharmacogenetic investigation of acute myeloid leukemia, we propose in our primary specific aim to systematically investigate the utility of using archived bone marrow samples for an investigation on the prognostic significance of a panel of genes involved in the pharmacodynamics of AML chemotherapy in a well-characterized group of AML patients. In our secondary specific aim, we propose to utilize the data generated from this methodological investigation for a pilot study on the role of this panel of polymorphisms relevant to AML treatment in toxicity and clinical outcome measures among AML patients. Specifically, we will compare genetic polymorphism data between paired bone marrow and buccal cells from 100 AML patients from Roswell Park Cancer Institute. In the pilot study component of this research we will assess role of genetic polymorphisms encoding for proteins involved in metabolism of chemotherapeutic agents used in the treatment of AML, protection from oxidative damage generated by chemotherapeutic agents, and drug resistance in clinical outcome measures among AML patients. Data generated from this research will guide the design of large-scale pharmacogenetic studies of AML by a) providing data on the appropriateness of using existing bone marrow tissue banks, and b) direct sample size considerations by providing data on potential misclassification of genotype data in bone marrow tissue, as well as preliminary data on the effect of genetic polymorphisms on clinical outcomes.